Digitalization facilitates insurance fraud - AI uncovers it

With its innumerable websites, blogs and forums, the Internet makes available an unimaginable amount of information. Thus, a search on Google also quickly leads to entries that explain the easiest way to submit a credible claim to an insurer.

As a representative survey commissioned by the GDV has shown, 6 percent of Germans admit to having reported damages to their insurance company that did not occur at all.[1] And 4 percent know someone who has done this. The good news is without doubt that the overwhelming majority of the insured are honest. Nevertheless, fraud remains a massive problem for the insurance industry.

Digital tools make evidence easier to fake

Long gone are the days when the allegedly injured parties submitted photographs in paper form to document a claim. Today, digital images are exchanged with the companies. To manipulate images with modern photo editors, no special knowledge is necessary anymore. Or photos available on the Internet are used to document damages. Claims that are covered by so-called cyber-insurances pose an even greater challenge because in this case, the alleged perpetrators and the alleged damages are almost exclusively in the sphere of bits & bytes. In short: Insurance fraudsters are upgrading digitally.

The Corona crisis could increase fraud figures

The pandemic is leading to a strained financial situation in many companies and private households. The money for the installments of a purchase could become scarce, or the owner of the company could have overextended himself with the purchase of goods. Thus, insurers fear that the number of “occasional fraudsters” could rise: A fictitious burglary, in which the (too) expensive television set or the no longer marketable fashion collection is stolen, ensures that the financial problems will be solved by the settlement of claims by the insurance company.

AI as an aid in the settlement of claims

It is not an easy task for employees in the claims department of insurance companies to detect fraudulent reports. Here, AI systems can effectively support case processing. Computer programs recognize faster and more precisely than humans whether a photo has been manipulated. Often, the meta-information stored in each image file already provides hints. With AI, the plausibility check becomes easier. Especially in the volume business of property or car insurance, self-learning systems recognize emerging patterns. AI uncovers new tricks faster than humans could, because information is linked much faster.

Investment in AI facilitates and accelerates case processing, saves costs and is also in the interest of the insured themselves. After all, it is their premiums that the fraudsters cash in.

[1] https://www.gdv.de/de/medien/aktuell/sorge-der-versicherer–corona-gibt-betruegern-auftrieb-61842